[parisc-linux] Re: Help on install

[parisc-linux] Re: Help on install

[Reinhard Gimbel]

> I'm attempting to to install HP Linux on a C180XP
> Has anyone had success install V0.9 on this type
> of machine?

I'm still trying to install 0.9.2 on a C180, too.
 
> The problem I'm having is that when booting from
> a CD, created with the current CD image on the
> web site, the system just hangs in the boot
> process after Serial driver has been enabled.

My expirence is that the DINO I/O chip seems to be the key 
element for that problem. Version 1.x of this chip crashes the 
system like you describe it (crash during boot ... :-( ) My 
DINO version (2.1) allows at least boot the system but the 
system crashes when I try to do the next steps within the 
installation procedure.

One tip i got in this mailing list was to try to install over 
network / NFS.

My problem at present is that the TP port of my C180 box seems 
to be broken, because i can't get a connection to my Intel 
Linux box. I tried to use either 100 Mbps as well as 10 Mbps on 
my Intel Linux box. The AUI-Ethernet port I can't test because 
I've not glue box from my Intel Linux TP ethernet adaptor to 
the AUI ethernet port of my C180 ...

Anybody outthere with an idea in this area ?

Thanks a lot !

Reinhard.

Re: [parisc-linux] Re: Help on install

[Matthew Wilcox]

On Thu, Jul 26, 2001 at 06:48:42PM +0200, Reinhard Gimbel wrote:
> My expirence is that the DINO I/O chip seems to be the key 
> element for that problem. Version 1.x of this chip crashes the 
> system like you describe it (crash during boot ... :-( )

Oooh.  That was the vital clue, see this note from the source:

** "HP has discovered a potential system defect that can affect
** the behavior of five models of HP VISUALIZE workstations when
** equipped with third-party or customer-installed PCI I/O expansion
** cards. The defect is limited to the HP C180, C160, C160L, B160L,
** and B132L VISUALIZE workstations, and will only be encountered
** when data is transmitted through PCI I/O expansion cards on the
** PCI bus. HP-supplied graphics cards that utilize the PCI bus are
** not affected."

Maybe we're triggering this bug... grant?


-- 
Revolutions do not require corporate support.

Re: [parisc-linux] Re: Help on install

[Grant Grundler]

Matthew Wilcox wrote:
> ** "HP has discovered a potential system defect that can affect
> ** the behavior of five models of HP VISUALIZE workstations when
> ** equipped with third-party or customer-installed PCI I/O expansion
> ** cards. The defect is limited to the HP C180, C160, C160L, B160L,
> ** and B132L VISUALIZE workstations, and will only be encountered
> ** when data is transmitted through PCI I/O expansion cards on the
> ** PCI bus. HP-supplied graphics cards that utilize the PCI bus are
> ** not affected."
> 
> Maybe we're triggering this bug... grant?

Perhaps...updating the firmware to the latest will not let the system
boot if in fact that's the case. remove all PCI devices and then
the system will boot.

grant

Grant Grundler
parisc-linux {PCI|IOMMU|SMP} hacker
+1.408.447.7253

[parisc-linux] C180 DINO/PCI problem; was: Help on install (1)

[Reinhard Gimbel]

Grant Grundler wrote:

> Matthew Wilcox wrote:
> > ** "HP has discovered a potential system defect that can affect
> > ** the behavior of five models of HP VISUALIZE workstations when
> > ** equipped with third-party or customer-installed PCI I/O expansion
> > ** cards. The defect is limited to the HP C180, C160, C160L, B160L,
> > ** and B132L VISUALIZE workstations, and will only be encountered
> > ** when data is transmitted through PCI I/O expansion cards on the
> > ** PCI bus. HP-supplied graphics cards that utilize the PCI bus are
> > ** not affected."

Expect the GRAPHICS(3) graphics adaptor (which is an EISA card)
there is no expansion card installed in my C180. The 2 PCI slots
are not occupied by expansion cards.

I don't have too much knowledge of the architecture of HP
workstations, but aren't NCR SCSI chips part of the PCI
architecture in ordinary PCs ? Of course there must not be the
same way to implement the SCSI chip in the HP architecture ! But
who knowns ...

On the other hand the C180 had a working HP-UX 10.20 as I bought
it weeks ago. So there must be a software solution for the DINO
problem.

> > Maybe we're triggering this bug... grant?
>
> Perhaps...updating the firmware to the latest will not let
> the system boot if in fact that's the case. remove all PCI
> devices and then the system will boot.

No PCI device to remove (as an expansion card) ...

... but how about a firmware update ? My box has firmware
version 2.7. Is that o.k. so far ? Or is there a newer firmware
available ? The other questions is: Would a firmware upgrade
help to solve my problem ?

Reinhard.

[parisc-linux] Re: C180 DINO/PCI problem; was: Help on install (1)

[Grant Grundler]

Reinhard Gimbel wrote:
> Expect the GRAPHICS(3) graphics adaptor (which is an EISA card)
> there is no expansion card installed in my C180. The 2 PCI slots
> are not occupied by expansion cards.
...
> On the other hand the C180 had a working HP-UX 10.20 as I bought
> it weeks ago. So there must be a software solution for the DINO
> problem.

Trust me. There isn't. On some of the boxes a HW workarond is possible.
But I'm not clear on if that ever got pushed out in replacement motherboards
to customers. I was under the impression the problem was "fixed" by
swapping for a motherboard with dino 3.1.

> > Perhaps...updating the firmware to the latest will not let
> > the system boot if in fact that's the case. remove all PCI
> > devices and then the system will boot.
> 
> No PCI device to remove (as an expansion card) ...

Ok.

> ... but how about a firmware update ? My box has firmware
> version 2.7. Is that o.k. so far? Or is there a newer firmware
> available ? The other questions is: Would a firmware upgrade
> help to solve my problem?

Definitely update. You want to know if your box is useable or not.

The FAQ has a nice section on firmware updates which I think
answers this set of questions.

grant

Grant Grundler
parisc-linux {PCI|IOMMU|SMP} hacker
+1.408.447.7253

Re: [parisc-linux] Re: Help on install

[Grant Grundler]

Reinhard Gimbel wrote:
> My expirence is that the DINO I/O chip seems to be the key 
> element for that problem. Version 1.x of this chip crashes the 
> system like you describe it (crash during boot ... :-( )

You have Dino 1.x ? I didn't think HP ever shipped those to customers.
Or did you mean another chip?
Perhaps post the console log output?

> My 
> DINO version (2.1) allows at least boot the system but the 
> system crashes when I try to do the next steps within the 
> installation procedure.

On C180 that's most likely the problem with ccio driver.
That's a known bug(s?)  that ryan bradetich hasn't been able to solve yet.
If you can help ryan, mail him.

grant

Grant Grundler
parisc-linux {PCI|IOMMU|SMP} hacker
+1.408.447.7253

[parisc-linux] C180 DINO/PCI problem; was: Help on install (2)

[Reinhard Gimbel]

Grant Grundler wrote:

> Reinhard Gimbel wrote:
> > My expirence is that the DINO I/O chip seems to be the key
> > element for that problem. Version 1.x of this chip crashes
> > the system like you describe it (crash during boot ... :-(
> > )
>
> You have Dino 1.x ? I didn't think HP ever shipped those to
> customers. Or did you mean another chip?
> Perhaps post the console log output?

A friend of mine has a C180 with DINO v1.x (according to the
boot messages of that box; I do not have a log of that box on 
hand at present).

The boot messages of my box (as posted in parisc-linux digest 
vol. 1 #204) looks like this:

[...]

Dino version 2.1 (bridge mode) found at 0xf2000000
 
 
The GSCtoPCI (Dino hrev 1) bus converter found may exhibit
data corruption.  See Service Note Numbers: A4190A-01, A4191A-01.
Systems shipped after Aug 20, 1997 will not exhibit this problem.
Models affected: C180, C160, C160L, B160L, and B132L workstations.
 
dino_bridge_init: IO_ADDR_EN hasn't been configured.
kernel BUG at dino.c:808!

[...]

     YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001101111100100001110
r0-3     00000000 102b2810 101c97cc 1f7a3a00
r4-7     1f7a3a28 00000006 00000000 00000000
r8-11    1f7a2000 000000ff 1fe02a74 00000200
r12-15   102b2810 00000001 00000400 1f584688
r16-19   1f584588 00000000 00000000 dae998dc
r20-23   00000008 1e1fc500 00000000 0000000f
r24-27   00000000 00000006 1f7a2000 1028a010
r28-31   00000000 00000400 1f584dc0 101bd938
sr0-3    00000000 000000b5 00000000 000000b5
sr4-7    00000000 00000000 00000000 00000000
 
IASQ: 00000000 00000000 IAOQ: 101d3004 101d3008
 IIR: 0e641288    ISR: 00000000  IOR: dae998e0
 ORIG_R28: 1073e740

[...]


> > My
> > DINO version (2.1) allows at least boot the system but the
> > system crashes when I try to do the next steps within the
> > installation procedure.
>
> On C180 that's most likely the problem with ccio driver.
> That's a known bug(s?)  that ryan bradetich hasn't been able
> to solve yet. If you can help ryan, mail him.

Ryan, what further information do you need ? Up to now I tried 
to use the palinux-0.9.2 image of esiee.fr only. No kernel 
(cross-) compilations for parisc-linux on my Intel Linux box so 
far. No expirence with palo etc. What can I do to help to solve 
the problem ?

Reinhard.

[parisc-linux] Re: C180 DINO/PCI problem; was: Help on install (2)

[Grant Grundler]

Reinhard Gimbel wrote:
> > You have Dino 1.x ? I didn't think HP ever shipped those to
> > customers. Or did you mean another chip?
> > Perhaps post the console log output?
> 
> A friend of mine has a C180 with DINO v1.x (according to the
> boot messages of that box; I do not have a log of that box on 
> hand at present).

Ok. IIRC, early revs of dino mostly had problems with nested PCI-PCI
bridges. I'd need to review the dino errata and I don't have that with
me here.

> Dino version 2.1 (bridge mode) found at 0xf2000000
...
> dino_bridge_init: IO_ADDR_EN hasn't been configured.
> kernel BUG at dino.c:808!

hmm...the newer PDC isn't programming Dino or anything
below it since the PDC knows it's buggy. Seems like dino code
should just give up trying to walk the bus at this point.
I suspect that's crashing the box later.

grant

Grant Grundler
parisc-linux {PCI|IOMMU|SMP} hacker
+1.408.447.7253